KR100946707B1 - Method, system and computer-readable recording medium for image matching of panoramic images - Google Patents

Abstract

PURPOSE: An image matching method of panoramic images, a system thereof, and a computer-readable recording medium thereof are provided to innovatively improve the accuracy of image matching between a panoramic image and a planar image. CONSTITUTION: An image matching method of panoramic images comprises the following steps: a step of classifying a panoramic image into multiple curve area; a step of generating an planar image by changing the multiple curve area to be displayed on a planar area depending on cartesian coordinate system; a step of extracting at least one characteristic area in the planar image; a step of normalizing at least one characteristic area; a step of deciding whether the planar image and a query image is similar if the query image is inputted.

Description

METHOD, SYSTEM AND COMPUTER-READABLE RECORDING MEDIUM FOR IMAGE MATCHING OF PANORAMIC IMAGES}

The present invention relates to a method, a system and a computer readable recording medium for performing image matching on a panoramic image. More specifically, the present invention converts a panoramic image into a planar image according to Cartesian coordinates in performing image matching between the panoramic image and the query image, and changes the viewpoint or lighting of the converted planar image. The present invention relates to a method, a system and a computer-readable recording medium for improving the accuracy of image matching between a panoramic image and a query image by reflecting and then matching the query image with the image.

In recent years, as the use of the Internet has spread, it has become common that geographical information, which has been provided in the form of printed books, is converted into digital data and provided. As a typical form of geographic information that is digitized, for example, an electronic map service provided online, a navigation system installed in an automobile, and the like. The digitized geographic information as described above has the advantage of being able to provide various convenience functions to the user by being combined with a user interface such as a search function, and when a change is made to the geographic information, it can be easily updated through a remote update. Therefore, the situation is superior to the conventional printed matter in terms of the latest geographic information.

However, in general, service providers who provide digital geographic information generally provide an electronic map screen generated by a simple combination of symbols in combination with a search interface. There is a difference, it is difficult to see the actual appearance of the area.

In an attempt to solve such a problem, US Patent Publication No. 2008-143727, filed on November 8, 2007, and published on June 19, 2008, provides a user with an electronic map service. Providing panoramic image data corresponding to a specific point such as an alley or a road appearing on an electronic map, and when a user selects a part of the panorama image data represented by an arrow or the like, corresponding to the point indicated by the corresponding pixel Another technique for providing panoramic image data is proposed. Accordingly, the user can confirm the actual state of a specific point on the electronic map through the panoramic image data, and the geographic information of the specific point indicated in the panoramic image data currently being viewed through the link generated between the panoramic image data is expressed in more detail. Other panoramic image data can be queried.

In addition, a technique of performing a visual search on a panoramic image has been introduced. According to this, an object included in the query image among a plurality of panoramic images with reference to a query image input by a user is introduced. A panoramic image containing similar objects can be retrieved and provided to the user.

However, since the image input as the query image is mainly a planar image constructed according to the Cartesian coordinate system, the panorama image matching this query image is a curved image composed according to the spherical coordinate or the cylindrical coordinate system. There is a problem in that the accuracy of the image matching is lowered when matching directly. Hereinafter, the problems of the conventional image matching technique will be described in detail with reference to FIGS. 1 to 4.

FIG. 1 is a diagram illustrating a query image, which is a planar image according to a Cartesian coordinate system, and FIG. 2 illustrates a feature region extracted from the query image of FIG. 1 (referring to a predetermined region around a feature point, also referred to as a local feature). It is a figure shown normally. 3 is a diagram illustrating a panoramic image according to a spherical coordinate system, and FIG. 4 is a diagram illustrating a feature region extracted from the panoramic image of FIG. 3.

First, referring to FIGS. 1 and 2, when an image such as FIG. 1 along a Cartesian coordinate system is inputted as a query, a feature region for an arrow indicating a traveling direction that exists in the sign portion 110 of FIG. 1 is illustrated in FIG. 1. As shown in 2, the normal extraction can be seen.

On the other hand, referring to Figures 3 and 4, in the case of the panoramic image as shown in Figure 3 along the spherical coordinate system, the feature region for the arrow, etc. present in the sign portion 310 of Figure 3 is properly extracted as shown in FIG. You can see that it is not. In general, since similarity determination between images is performed by comparing feature points and feature regions extracted from each image, when the feature regions of the panoramic image are not properly extracted as shown in FIGS. Problems arise in that accurate image matching between planar images (ie, images according to the Cartesian coordinate system) and panoramic images becomes difficult.

The object of the present invention is to solve all the above-mentioned problems.

The present invention also converts a panoramic image into a planar image according to a Cartesian coordinate system, reflects a viewpoint or lighting change with respect to the converted planar image, and then performs image matching on the basis of the planar image, thereby providing an image for the panoramic image. Another purpose is to allow matching to be performed more accurately.

Representative configuration of the present invention for achieving the above object is as follows.

According to an aspect of the present invention, there is provided a method for performing image matching on a panoramic image, the method comprising: (a) dividing a panoramic image into a plurality of curved areas; and (b) dividing the plurality of curved areas into a Cartesian coordinate system. (c) extracting at least one feature region centered on at least one feature point on the plane image, (d) extracting the plane image by transforming the image to be displayed on the planar region according to coordinates; Normalizing one feature region, and (e) if a query image is input, matching the normalized feature region of the planar image with the normalized feature region of the query image to determine whether there is a similarity between the planar image and the query image A method is provided that includes determining.

According to another aspect of the present invention, there is provided a system for performing image matching on a panoramic image, wherein the panoramic image is divided into a plurality of curved areas, and the plurality of curved areas are arranged on a planar area according to a Cartesian coordinate. An image conversion unit for generating a planar image by converting the image to be displayed on a feature; And a query execution unit that determines whether the planar image and the query image are similar by matching the normalized feature region of the planar image with the normalized feature region of the query image when the query image is input. do.

In addition, there is further provided a computer readable recording medium for recording another method, system for implementing the present invention and a computer program for executing the method.

According to the present invention, in performing image matching between a panoramic image according to a spherical coordinate system or a cylindrical coordinate system and a planar image according to a Cartesian coordinate system, the panoramic image is converted into a planar image and a change in viewpoint or illumination is performed on the converted planar image. Applying the reflected transform matrix and matching between planar images allows feature images and feature regions of the panoramic image to be extracted more accurately, dramatically improving the accuracy of image matching between panoramic and planar images Can achieve the effect.

DETAILED DESCRIPTION The following detailed description of the invention refers to the accompanying drawings that show, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different but need not be mutually exclusive. For example, certain shapes, structures, and characteristics described herein may be embodied in other embodiments without departing from the spirit and scope of the invention with respect to one embodiment. In addition, it is to be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. The following detailed description, therefore, is not to be taken in a limiting sense, and the scope of the present invention, if properly described, is defined only by the appended claims, along with the full range of equivalents to which such claims are entitled. Like reference numerals in the drawings refer to the same or similar functions throughout the several aspects.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement the present invention.

[Preferred Embodiments of the Invention]

In the present specification, a panoramic image refers to an image obtained by photographing a complete view viewed at an arbitrary point, and more specifically, pixels constituting the image are centered on the photographing point. It is a generic term for a type of image that can be displayed according to spherical coordinates on a virtual celestial sphere, so that visual information on all four directions actually seen at a shooting point can be provided in three dimensions and realistically. will be. In addition, although not directly illustrated herein, it is noted that the panoramic image may be an image in which pixels constituting the image are displayed according to cylindrical coordinates or a similar coordinate system.

Configuration of the entire system

5 is a diagram schematically illustrating a configuration of an entire system for providing an image matching service for a panoramic image according to an embodiment of the present invention.

As shown in FIG. 5, the entire system according to an embodiment of the present invention includes a communication network 100, an image matching system 200 that provides an image matching service for a panoramic image, and a user terminal device 300. Can be configured.

First, the communication network 100 may be configured regardless of its communication mode such as wired and wireless, and may include a local area network (LAN), a metropolitan area network (MAN), and a wide area network (WAN). Network). Preferably, the communication network 100 according to the present invention may be a known World Wide Web (WWW).

According to an embodiment of the present invention, in providing a visual search service using a panoramic image to the user terminal device 300, the image matching system 200 divides the panoramic image into a plurality of curved areas and divides the plurality of curved areas. A function of generating a plane image may be performed by converting the plane to be displayed on a plane area according to a Cartesian coordinate. In addition, the image matching system 200 according to an embodiment of the present invention may extract and normalize at least one feature region centered on at least one feature point on the converted planar image, and the query image may be input. When the feature point and the normalized feature region of the planar image are matched with the feature point and the normalized feature region of the query image, a function of determining whether the plane image is similar to the query image may be performed. A detailed description of the internal configuration of the image matching system 200 will be described later.

On the other hand, the user terminal device 300 according to an embodiment of the present invention is a digital device that includes a function that allows the user to communicate after connecting to the image matching system 200, a personal computer (for example, desktop Computer, laptop computer, etc.), a workstation, a PDA, a web pad, a mobile phone, etc., any memory having a memory means and equipped with a microprocessor equipped with arithmetic capabilities, any number of the digital terminal adopting as the user terminal device 300 according to the present invention Can be.

Configuration of Image Matching System

Hereinafter, the internal structure of the image matching system 200 performing important functions for the implementation of the present invention and the functions of each component will be described.

6 is a diagram illustrating an internal configuration of an image matching system 200 according to an embodiment of the present invention.

Referring to FIG. 6, the image matching system 200 according to an embodiment of the present invention may include an image converter 210, a feature extractor 220, a normalizer 230, a query performer 240, and a database. 250, a communication unit 260, and a control unit 270 may be included. According to an embodiment of the present invention, the image converter 210, the feature extractor 220, the normalizer 230, the query execution unit 240, the database 250, the communication unit 260 and the control unit 270 At least some of them may be program modules that communicate with the user terminal device 300. Such program modules may be included in the image matching system 200 in the form of operating systems, application modules, and other program modules, and may be physically stored on various known storage devices. In addition, these program modules may be stored in a remote storage device that can communicate with the image matching system 200. On the other hand, such program modules include, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc. that perform particular tasks or execute particular abstract data types, described below, in accordance with the present invention.

First, according to an embodiment of the present invention, the image converting unit 210 divides the panoramic image according to the spherical coordinate system or the cylindrical coordinate system, etc. into a plurality of curved areas, and the divided plurality of curved areas are Cartesian coordinates. A function of generating a planar image by converting the image to be displayed on a planar region according to FIG. According to an embodiment of the present invention, the at least one feature point extracted in the panoramic image may be included in each curved area so that the main feature points may be expressed intactly in the converted planar image, but the present invention is not limited thereto. In some cases, various modifications may be envisioned, such that the feature point may not be included in a specific curved area, or may be performed after converting the feature extraction operation into a planar image.

More specifically, the image conversion unit 210 according to an embodiment of the present invention, by mapping the pixel value of the pixel included in each divided curved area of the panoramic image by reference to a predetermined plane area, The conversion to the planar image is performed. An algorithm for overcoming the difference between the number of pixels included in the curved area and the number of pixels included in the planar area in the mapping operation will be described below with reference to FIGS. 7 to 9.

7 to 9 are conceptual views illustrating a principle of converting a panoramic image into a planar image according to an embodiment of the present invention.

More specifically, the image conversion unit 210 according to an embodiment of the present invention, by mapping the pixel value of the pixel included in each divided curved area of the panoramic image by reference to a predetermined plane area, The conversion to the planar image is performed. An algorithm for overcoming the difference between the number of pixels included in the curved area and the number of pixels included in the planar area in the mapping operation will be described below with reference to FIGS. 7 to 9.

7 to 9 are conceptual views illustrating a principle of converting a panoramic image into a planar image according to an embodiment of the present invention.

In addition, referring to FIG. 7, the planar image 530 converted from the plurality of curved areas 520 of the panoramic image extends from the center of the sphere on which the panoramic image 510 is projected, respectively, according to a predetermined latitude and longitude. Four semi-linear lines may be formed as rectangular facets having four vertices each of which meets the sphere.

In addition, according to an embodiment of the present invention, the pixel value of each pixel constituting the planar image 530 refers to the pixel value of the pixel included in the curved area 520 corresponding to the planar image 530 of the panoramic image. It can be determined by.

More specifically, the image converting unit 210 according to an exemplary embodiment of the present invention projects a region occupied by a specific pixel on the planar image 530 onto the curved surface area 520, and thus, on the curved surface area 520. Comparing the projection area to be formed and the area occupied by each pixel of the curved area 520, and referring to pixel values of at least one pixel among the pixels of the curved area 520 covered by the projection area. The pixel value of the specific pixel included in the raw plane image 530 may be determined. Here, the pixel value of the specific pixel on the planar image 530 is a weighted average of pixel values of at least one pixel of at least some of the pixels of the curved area 520 covered by the projection area. May be determined according to the area of the overlapping area of each pixel on the projection area and the corresponding curved area 520.

Meanwhile, according to an embodiment of the present invention, the image converter 210 may estimate the latitude and longitude of a specific pixel on the planar image by applying interpolation with reference to the latitude and longitude of the four intersections. There will be.

Next, according to an embodiment of the present invention, the feature extractor 220 extracts the feature point, which is a reference of image matching, from the panoramic image before being converted by the image converter 210, and the image converter ( The feature region may be extracted from the planar image generated by 210, but is not necessarily limited thereto. For example, the feature point and the feature region may be extracted from the planar image generated by the image converter 210. You can also extract Here, the feature point refers to a point including the feature elements of the object included in the image, and the feature area refers to the area around the feature point including the object's characteristics, so that the feature point is robust to the view point and the light change of the image. Can be set.

Meanwhile, as mentioned above, in order to extract feature points and feature regions from an image, a predetermined feature extraction technique is required. According to one embodiment of the present invention, as a feature recognition technique, C. Harris et al. It may be referred to a paper entitled "A combined corner and edge detector" described in "In Alvey Vision Conference", 1988 (the contents of which should be considered to be incorporated in its entirety). . This paper describes a method for estimating an elliptic feature region using a second moment matrix representing a distribution of slopes around a feature point. Of course, the object recognition technology applicable to the present invention is not limited to the method described in the above paper, and various modifications may be applied to implement the present invention.

On the other hand, since the query image and the planar image that are the target of image matching are images captured through different viewpoints in different lighting environments, they are included in the planar image converted from the panoramic image to follow the query image and the Cartesian coordinate system. The feature region may be extracted differently in size and shape according to a viewpoint or lighting. As a result, it is difficult to perform accurate image matching only by directly comparing the feature region of the query image with the feature region of the transformed planar image.

Accordingly, the normalization unit 230 according to an embodiment of the present invention solves the above problem caused by different sizes and shapes of feature regions included in the query image and the transformed planar image, which are the targets of image matching, respectively. In order to solve the problem, a field of view between the query image and the transformed planar image is normalized by normalizing the size and shape of the feature region included in the query image and the transformed planar image, and performing image matching based on the normalized feature region. The error caused by the difference in illumination can be corrected.

10 is a diagram exemplarily illustrating a configuration for normalizing a feature region according to an embodiment of the present invention. Image (a) of FIG. 10 shows feature area A 810 extracted from the query image as part of the query image, and image (b) of FIG. 10 shows feature area extracted from the planar image as part of the converted planar image. B 820 is shown. Looking at the images (a) and (b) of FIG. 10, the query image and the transformed planar image, due to a viewpoint or lighting difference, actually represent the same object, but the same feature parts of the same object have different sizes and shapes in each image. As can be seen, it can be seen that feature regions 810 and 820 having different sizes and shapes are extracted from the same feature portion of the query image and the converted planar image.

The normalization unit 230 according to an embodiment of the present invention may normalize a pair of feature regions having different sizes and shapes into a pair of feature regions having the same size and shape using a predetermined normalization technique. Accordingly, the pair of feature areas 810 and 820 as shown in the images (a) and (b) of FIG. 10 may correspond to the pair of feature areas 830 as shown in the images (c) and (d) of FIG. 10. 840).

Meanwhile, as mentioned above, in order to normalize the feature region extracted from the image, a predetermined specific region normalization technique is required. According to an embodiment of the present invention, as such feature region normalization technique, K. MIKOLAJCZYK et al. The author may refer to a paper entitled "A Comparison of Affine Region Detectors" in the November 2005 "International Journal of Computer Vision" (the content of which is incorporated herein in its entirety). Should be considered). In this paper, by using M _L ^1/2 and M _R ^1/2 as the second moment matrix for estimating the viewpoint and lighting conditions of the image, the feature region composed of ellipses having various sizes and shapes has a specific size and shape. A method of normalizing a circle is described, and a method of rotating a normalized feature region using a rotation matrix R to determine whether a pair of normalized feature regions point to the same object is described. Of course, the normalization technique applicable to the present invention is not limited to the method described in the above paper, and various modifications may be applied to implement the present invention.

According to an embodiment of the present invention, the query execution unit 240 may perform image matching to determine the similarity between the query image and the transformed planar image based on the normalization method of the feature region described above. More specifically, the query execution unit 240 according to an embodiment of the present invention compares at least one normalized feature region of the query image with at least one normalized feature region of the transformed planar image, and thus, At least one pair of feature regions determined to point to the same object may be searched for. In addition, when at least two pairs of feature areas are searched, the query execution unit 240 according to an embodiment of the present invention uses at least two image features corresponding to the at least two pairs of feature areas using a topology. Compare the relative positional relationship between two feature regions and the relative positional relationship between at least two feature regions on the transformed planar image corresponding to the at least two pairs of feature regions, and refer to the query image and the plane with reference to the comparison result. The similarity of the image can be determined.

FIG. 11 is a diagram exemplarily illustrating a distribution of feature areas included in a query image and a converted planar image, according to an exemplary embodiment. Referring to FIG. 11, the query execution unit 240 according to an embodiment of the present invention may include a relative positional relationship between a plurality of feature areas on a query image (ie, image (a)) and a transformed plane image (ie, an image ( By comparing the relative positional relationship of the plurality of feature areas on b)), it is possible to determine whether the query image and the transformed planar image are similar.

On the other hand, as mentioned above, in order to determine the similarity between two different images by using the relative positional relationship of the feature region, a predetermined topology technique is required. According to one embodiment of the present invention, , By DERDAR Salah and two others, may refer to a paper entitled "Image matching using algebraic topology" in January 2006 "Proceedings of SPIE, Vol. 6066". Should be considered to be incorporated herein). This paper describes a method for performing image matching by measuring the similarity between boundary elements of features included in an image using an algebraic topology. Of course, the topology technology applicable to the present invention is not limited to the method described in the above paper, and various modifications may be applied to implement the present invention.

Further, according to an embodiment of the present invention, if it is determined that the query image is similar to the deformed plane image, the query execution unit 240 searches for a query image for a portion corresponding to the deformed plane image of the panorama image. The information about the object included in the query image may be obtained from a portion corresponding to the modified planar image among the matched panoramic images, and provided.

As described above, according to the image matching method according to the present invention, an effect of improving the accuracy of image matching between the panoramic image and the query image is achieved.

In particular, the image matching method of removing curved surface characteristics of a panoramic image and correcting an error according to a viewpoint and lighting change according to the present invention is a conventional method of correcting only an error according to a viewpoint and lighting change without considering characteristics of a panoramic image. Compared to the image matching method of can provide a more accurate image matching results.

12 is a diagram illustrating a feature region extracted from a panoramic image according to a conventional image matching method, and FIG. 13 is a diagram illustrating a feature region extracted from a panoramic image according to an image matching method of the present invention. .

Referring to FIG. 12, according to the conventional image matching method, even though the panoramic image has curved surface characteristics by directly extracting the feature region 1210 from the panoramic image according to the spherical coordinate system without considering the characteristics of the panoramic image. Since a specific region comes out in the shape of a smooth oval, the extracted feature region 1210 may have a large difference from the feature region (not shown) extracted from the query image consisting of a planar image according to the Cartesian coordinate system. This may cause the accuracy of image matching to be lowered.

On the other hand, referring to Figure 13, according to the image matching method according to the present invention, after converting the panoramic image according to the spherical coordinate system into a planar image according to the Cartesian coordinate system and reflecting the viewpoint or lighting changes in the converted planar image Since the region 1310 is extracted, the extracted feature region 1310 may correspond more accurately to the feature region (not shown) extracted from the query image composed of the planar image, thereby improving the accuracy of matching. Can be. Meanwhile, considering the difference between the panoramic image displayed on the curved surface and the planar image displayed on the plane, the feature region 1310 extracted from the planar image according to the image matching method of the present invention has a shape of an ellipse that is somewhat bent on the panoramic image. It can be expressed as

On the other hand, in the present invention, the database 250 is a concept that includes not only a negotiated database but also a database of a broad meaning including a data record based on a computer file system, and the like. It should be understood that the data of the present invention can be included in the database of the present invention if it can be extracted. Although the database 250 is illustrated as being included in the image matching system 200 in FIG. 6, according to the needs of those skilled in the art of implementing the present invention, the database 250 may be connected to the image matching system 200. It may be configured separately.

Meanwhile, the communication unit 260 according to an embodiment of the present invention performs a function of allowing the image matching system 200 to communicate with an external device such as the user terminal device 300.

In addition, the control unit 270 according to an embodiment of the present invention is an image converter 210, feature extraction unit 220, normalization unit 230, query execution unit 240, database 250 and communication unit 260 It controls the flow of data between). That is, the controller 270 controls the flow of data from the outside or between the respective components of the image matching system 200, such that the image converter 210, the feature extractor 220, the query performer 240, The database 250 and the communication unit 260 controls to perform a unique function, respectively.

Embodiments according to the present invention described above may be implemented in the form of program instructions that may be executed by various computer components, and may be recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the computer-readable recording medium may be those specially designed and configured for the present invention, or may be known and available to those skilled in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROMs, DVDs, and magneto-optical media such as floptical disks. media), and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform the process according to the invention, and vice versa.

Although the present invention has been described by specific embodiments such as specific components and the like, but the embodiments and the drawings are provided to assist in a more general understanding of the present invention, the present invention is not limited to the above embodiments. For those skilled in the art, various modifications and variations can be made from these descriptions.

Accordingly, the spirit of the present invention should not be limited to the above-described embodiments, and all of the equivalents or equivalents of the claims, as well as the appended claims, fall within the scope of the spirit of the present invention. I will say.

FIG. 1 is a diagram illustrating a query image that is a planar image according to the Cartesian coordinate system, and FIG. 2 is a diagram illustrating a feature region extracted from the query image of FIG. 1. 3 is a diagram illustrating a panoramic image according to a spherical coordinate system, and FIG. 4 is a diagram illustrating a feature region extracted from the panoramic image of FIG. 3.

5 is a diagram schematically illustrating a configuration of an entire system for providing an image matching service for a panoramic image according to an embodiment of the present invention.

6 is a diagram illustrating an internal configuration of an image matching system 200 according to an embodiment of the present invention.

7 to 9 are conceptual views illustrating a principle of converting a panoramic image into a planar image according to an embodiment of the present invention.

10 is a diagram exemplarily illustrating a configuration for normalizing a feature region according to an embodiment of the present invention.

FIG. 11 is a diagram exemplarily illustrating a distribution of feature areas included in a query image and a planar image, according to an exemplary embodiment.

12 is a diagram illustrating a feature region extracted from a panoramic image according to a conventional image matching method, and FIG. 13 is a diagram illustrating a feature region extracted from a panoramic image according to an image matching method of the present invention.

<Brief description of the major reference numerals>

100: network

200: image matching system

210: image conversion unit

220: feature extraction unit

230: query execution unit

240: database

250: communication unit

260: control unit

300: user terminal device

Claims (21)

A method for performing image matching on a panoramic image, the method comprising: (a) dividing the panoramic image into a plurality of curved areas, (b) generating a planar image by converting the plurality of curved areas to be displayed on a planar area according to a Cartesian coordinate; (c) extracting each of at least one feature region centered on at least one feature point on the planar image; (d) normalizing the at least one feature region, and (e) if a query image is input, determining a similarity between the planar image and the query image by matching a normalized feature region of the planar image with a normalized feature region of the query image How to include. The method of claim 1, In the step (a), And extracting at least one feature point from the panoramic image and dividing the panoramic image into a plurality of curved areas based on the extracted position of the at least one feature point. The method of claim 1, In step (b), The planar image is composed of rectangular facets having four vertices each of which meets the sphere with four vertical lines extending from the center of the sphere on which the panoramic image is projected according to a predetermined latitude and longitude. How to. The method of claim 3, In step (b), The projection area formed on the sphere is compared with the area occupied by each pixel of the panoramic image as a result of projecting an area occupied by a specific pixel of the planar image on the plane image on the plane image, and among the pixels of the panoramic image. Determining a pixel value of the specific pixel of the planar image with reference to a pixel value of at least one pixel covered by at least a partial region by the projection area. The method of claim 4, wherein In step (b), The pixel value of the specific pixel of the planar image is a weighted average of pixel values of at least one pixel of which at least a portion of the pixel of the panoramic image is covered by the projection area, wherein the weight is the projection area and the panorama. The area of the overlapping area of each pixel on the image. The method of claim 3, In step (b), Estimating the latitude and longitude of a particular pixel on the planar image using interpolation with reference to the latitude and longitude of the four intersections. The method of claim 1, In step (e), (e1) comparing at least one normalized feature region of the planar image with at least one normalized feature region of the query image to search for at least a pair of feature regions determined to point to the same object as each other; One of the pair of feature regions is searched from the planar image and the other is searched from the query image; and (e2) when at least two pairs of feature regions are searched, a relative positional relationship between at least two feature regions on the planar image corresponding to the at least two pairs of feature regions using a topology and the at least two pairs Comparing a relative positional relationship between at least two feature regions on the query image corresponding to the feature region of the image, and determining whether the planar image matches the query image Method comprising a. The method of claim 7, wherein In the step (e1), Rotating said normalized feature region in performing said comparison. The method of claim 1, In the step (e), And said normalization is performed by converting the shape of said feature region from an elliptic form to a circular form using a transformation matrix. The method of claim 1, (f) if it is determined that the query image is similar to the planar image, providing a portion corresponding to the planar image or information about the planar image as a search result; Method further comprising a. A system for performing image matching on a panoramic image, the system comprising: An image converter for dividing the panoramic image into a plurality of curved areas, and converting the plurality of curved areas to be displayed on a planar area according to a Cartesian coordinate system to generate a planar image; A feature extractor configured to extract at least one feature region centered on at least one feature point on the planar image; A normalizer for normalizing the at least one feature region, and When a query image is input, a query execution unit that matches the normalized feature region of the planar image with the normalized feature region of the query image to determine whether the planar image is similar to the query image. System comprising a. The method of claim 11, The feature extraction unit, Extracting at least one feature point on the panoramic image, The image conversion unit, And the panoramic image is divided into a plurality of curved areas based on the extracted positions of the at least one feature point. The method of claim 11, The image conversion unit, A rectangular image having a vertex as four vertices each of which extends from the center of the sphere on which the panoramic image is projected according to a predetermined latitude and longitude respectively meets the sphere as a planar image; System. The method of claim 13, The image conversion unit, The projection area formed on the sphere is compared with the area occupied by each pixel of the panoramic image as a result of projecting an area occupied by a specific pixel of the planar image on the plane image on the plane image, and among the pixels of the panoramic image. And determine a pixel value of said particular pixel of said planar image with reference to a pixel value of at least one pixel covered by at least a portion of said projection area. The method of claim 14, The image conversion unit, A pixel value of the specific pixel of the planar image is a weighted average of pixel values of at least one pixel of which at least a portion of the pixel of the panoramic image is covered by the projection area, wherein the weight is the projection area and the panorama. System according to the area of the overlapping area of each pixel on the image. The method of claim 13, The image conversion unit, And estimate latitude and longitude of a particular pixel on the planar image by using interpolation with reference to the latitude and longitude of the four intersections. The method of claim 11, The normalization unit, Comparing at least one normalized feature region of the planar image with at least one normalized feature region of the query image to search for at least a pair of feature regions determined to point to the same object as each other-the pair of features One of the regions is searched from the planar image and the other is searched from the query image-when at least two pairs of feature areas are searched, the corresponding to the at least two pairs of feature areas using topology Compare the relative positional relationship between at least two feature regions on the planar image with the relative positional relationship between at least two feature regions on the query image corresponding to the at least two pairs of feature regions, and compare the planar image with the query image A system for determining whether to match. The method of claim 17, The normalization unit, And rotate the normalized feature region in comparing at least one normalized feature region of the planar image with at least one normalized feature region of the query image. The method of claim 11, The normalization unit, And performing normalization by converting the shape of the feature region from an ellipse to a circle using a transformation matrix. The method of claim 11, The query execution unit, And if it is determined that the query image is similar to the planar image, a portion corresponding to the planar image or information about the planar image is provided as a search result. A computer readable recording medium having recorded thereon a computer program for executing the method according to any one of claims 1 to 10.

Images